AMERICAN MUSEUM NOVITATES Published by Tnui Amermican MUSZUM W Number 632 Near York Cityratt1ral Historay June 9, 1933

Home , Bramatherium, Giraffokeryx, Palaeotragus, Protoceratidae

AMERICAN MUSEUM NOVITATES Published by Tnui AmERMICAN MUSZUM W Number 632 Near York CityRATt1RAL HisToRay June 9, 1933

56.9, 735 G: 14.71, 4 A SKULL AND MANDIBLE OF GIRAFFOKERYX PUNJABIENSIS PILGRIM By EDWIN H. COLBERT The genus Giraffokeryx was founded by Dr. G. E. Pilgrim to desig- nate a primitive Miocene giraffe from the lower Siwalik beds of northern India. Doctor Pilgrim, in a series of papers,' described Giraffokeryx on the basis of fragmental and scattered dentitions.. Naturally, Pilgrim's knowledge of the genus was rather incomplete, and he was unable tQ formulate any opinions as to the structure.of the skull or mandible. An almost complete skull, found in the northern Punjab in 1922 by Mr. Barnum Brown of the American Museum, proves to be that of Giraffokeryx, and it exhibits such striking and unusual characters that a separate description of it has seemed necessary. This skull, together with numerous teeth and a lower. jaw, gives us. a very good comprehen- sion of the genus which forms the subject.of this paper. The drawings of the skull were made by John. C. Germann, and the remaining ones were done by Margaret Matthew. MATERIAL DESCRIBED Only the material referred to in this description will here be listed. There' are a great many specimens of Gir'affokeryx in the American'Mu- seum collection, but since 'most of them are'teeth, they will not be considered at this time. A subsequent paper, dealing with the American Museum Siwalik collection in detail, wtyill contain a complete list of the Giraffokeryx material. Amer. Mus. No. 19475.-A skull, complete back of'the muzzle. The cheek teeth are preserved on both sides. The skull has been crushed to some extent. From the lower portion of the Middle Siwaliks, 1,000 feet below the Bhandar bone bed,2 and one mile south of Nathot in the Punjab. Amer. Mus. No. 19611.-Unassociated teeth. Lower Siwaliks, 200 feet above the level of Chinji Rest House and four miles west of that location, Punjab. lPilgrim, 0. E. 1910. 'Notices of New Mammalian Genera and Species from the Tertiaries of India.' Rec. Geol. Surv. India, XL, Pt. I, p. 69. 1911. 'The Fossil Giraffidae of India.' Mem. Geol. Surv. India, Pal. Indica, N.S., IV, No. 1, pp. 14-17; PI. 1, figs. 4, 5; P1. It, figs. 1-16. .2The Bhandar bed was named by Pilgrim, from the village of Bhandar in the Jhelum district, Rawalpindi division of the northern Punjab. He considered this 'as an unfossiliferous horizon, but Brown, in 19?22 opened a mixed quarry at this level, which yielded characteristic nmiddle Siwalik fossil. The quarry was located one half mile north of Bhandar village. For the original designation of'this horizon the reader is referred to: Pilgrim, G. E. 1913; 'Correlatiorf of the Siwaliks with Mammal Horizons of Europe. Rec. Geol. Surv. India, XLIII, Pt. 4, p. 277, P1. xxvi. -2 AMERICAN MUSEUM NOVITATES [No. 632

Amer. Mus. No. 19472.-A right maxilla, with P3-M3. Lower portion of the middle Siwaliks, 1,200 feet below the Bhandar bone bed and one mile south of Nathot. Amer. Mus. No. 19587.-The right ramus of a mandible, containing P2-M3. Lower Siwaliks, 200 feet above the level of Chinji Rest House and one and one half miles north of that place. It may be seen from the above list that Giraffokeryx punjabiensis extends through the lower Siwaliks into the lower portions of the middle Siwaliks. MORPHOLOGY OF THE SKULL The skull of Giraffokeryx represents an animal of medium size and of peculiar form. As restored, it would measure some 500 millimeters in length, which though of fair size for the general run of ruminants, is .small for the giraffes. The most striking features of this skull are the four horn cores, an anterior and a posterior pair, the former rising from the forward extremities of the frontals, and the latter from broad bases that rest partly on the frontals and partly on the parietals. The anterior *horn cores spring from a common base, which is located above the first and second molars (perhaps it might be better to say that they are confluent at their bases), and they diverge laterally at an angle of approximately 105 degrees. These horn cores are directed posteriorly to a very slight extent, about 15 degrees. The posterior horns are widely separated basally. They arise directly back of the orbits, diverging laterally at an angle of approximately 80 degrees. Moreover, these horn cores are directed backwardly to an extent such that, as seen from the side, their median axes make angles of about 30 degrees with the perpen- dicular. These posterior horn cores are somewhat convex along their anterior borders, and at their bases, both on the anterior and the posterior edges, there are large rugose protuberances, as if in life there were basal accessory knobs or branches. Due to the great backward growth of these horn cores, their bases form a partial roof over the temporal fossae. The true homologies of the "horn cores" of Giraffokeryx have been .very perplexing, because since the skull is that of an old animal, the sutures are ankylosed to an extent that they have been largely obliterated. By a microscopic examination the naso-frontal suture has been estab- lished just anterior to the base of the front horn cores. The frontal- parietal suture can not be definitely located, but it would seem to cross the skull at a point about opposite the median portion of the posterior horn cores. The posterior horn cores of Giraffokeryx are therefore quite homologous with the similar structures in Palaeotragus and Samotherium, Asiatic Miocene and Pliocene giraffes, characterized by one pair of well 19331 SKULL OF GIRAFFOKERYX PUNJABIENSIS 3 developed supraorbital horn cores. But the anterior pair in Giraffokeryx are without homologues among any of the Giraffidae, and must therefore be considered as neomorphic structures. Of course there is a superficial resemblance between the skull of Giraffokeryx and that of the North American genus, Syndyoceras, because in both of these forms there are two pairs of horns, one pair being approximately supraorbital, while the other pair is placed above the muzzle. The resemblances between these genera are, however, attribut- able to functional analogies rather than to homologies, because while the

Adt19475

Fig. 1.-Giraffokeryx punjabienris Pilgrim. Amer. Mus. No. 19475. Skull, right lateral view. Restored muzzle based on comparisons with Palaeotragus and Okapia. One-fourth natural size. anterior horns are frontal upgrowths in Giraffokeryx, they are derived in Syndyoceras from the premaxillaries. In both genera the posterior horns are more nearly homologous, being frontal in Syndyoceras, and fronto- parietal in Giraffokeryx. Otherwise the skulls and dentitions show the taxonomic separateness that would be expected in these genera, for the Siwalik form belongs of course to the Giraffidae, and the North American genus is a member of the Protoceratidae, families that are related in kind, but which are nevertheless quite distinct from each other. Synrhetoceras, recently described by Stirton, shows an accentuation of the Syndyoceras type of skull, and these two genera are closely related to each other.' iBarbour, E. H. 1905. 'Notice of a New Fossil Mammal from Sioux County, Nebraska.' Neb. Geol. Surv., II, P1. 3, four pages, one plate. (Syndyoceras cooki.) Stirton, R. A. 1932. 'A New Genus of Artiodactyla from the Clarendon Lower Pliocene of Texas.' Univ. of California Publ., Bull. Dept. Geol. Sci., XXI, No. 6, pp. 148-149. (Synthetoceras tricornatus.) AM.19475

Fig. ?.-Giraffokeryx punjabienis Pilgrim. Amer. Mus. No. 19475. Skull, dorsal view. Muzzle restored. One-fourth natural size.

4 1933] SKULL OF GIRAFFOKERYX PUNJABIENSIS 5 Lankester,1 in a series of papers, has demonstrated how in the modern okapi and giraffe, the "horns," or the ossicones as Lankester defined them, ate developed in the skin as separated elements, becoming united with the skull as the animal approaches maturity. In the okapi, a persistent primitive form, having, like Palaeotragus and Samotherium, one pair of supraorbital horns, the ossicone coalesces with the frontal bone, and in the adult animal is firmly attached to that element. On the other hand, the ossicone in the foetal or newly born giraffe develops first above the parietal bone, but in consequence of growth factors it spreads anteriorly, so that in the adult it becomes attached to both the parietal and the frontal bones. A growth process similar to that of the recent forms must have marked the development of the horns in Giraffokeryx. Indeed, the overlap of the base of each horn core on the roof of the skull is a striking character in this genus. Thus we may visualize small ossicones develop- ing in the skin of the young Giraffokeryx, one pair above the anterior border of the frontals, and another pair over the back of the frontals. At maturity the front pair became attached to the frontal bones, but the posterior ossicones spread into long, elliptical bases that overlapped both the frontals and the parietals. The nasals terminate posteriorly at the base of the front horn cores. The frontals are very wide, and extraordinarily flat, and behind them the parietals narrow rapidly and then flare out again to form the lambdoidal crest. The great constriction of the parietals in Giraffokeryx is an indication of the small size of the brain. The parietals dip down somewhat just behind the frontals, reaching the lowest point above the paroccipital processes, and then they rise again to the lambdoidal crest, which projects far behind the occipital condyles. There are supraQrbital foramina piercing the frontals. As seen from the side, the skull appears to be very low, an illusion caused by the crushing that it has undergone. By measuring the maxilla below the anterior horn cores, I have estimated the crushing to be about 17 per cent., so that when the skull is restored to its orignial height it is seen to be quite comparable to the skull of Palaeotragus. The basicranial axis forms a slight angle with the palatal surface. The maxilla is heavy, with a slight swelling above the first and second molars, and the malar runs far forward on this bone. The latter forms the entire lower border of the orbit. The lacrymal is large, a 1Lankester, E. Ray. 1907. 'The Origin of the Lateral Horns of the Giraffe in Foetal Life on the Area of the Parietal Bones.' Proc. Zool. Soc. London, pp. 100-115. 1907. 'The Existence of Rudimentary Antlers in the Okapi.' Proc. Zool. Soc. London, pp. 126-135. A.M.19475

/ 1, ;

I //

$t1k11

Fig. 3.-Giraffokeryx punjabienis Pilgrim. Amer. Mus. No. 19475. Skull, palatal view. Muzzle restored. One-fourth natural size.

6~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1933] SKULL OF GIRAFFOKERYX PUNJABIENSIS 7 portion of it being within the orbit, and a larger portion extending forward on the side of the face. Where the lacrymal, malar and jugal meet there is a large preorbital vacuity, a characteristic feature of the giraffids.. It would appear as if the trigeminus nerve traversed a maxillary sinus. cavity, as in the recent giraffes. The orbit is large, and is closed by a postorbital bar. Behind the orbit the large base of the posterior horn core projects out and back, so that it forms a heavy shelf or roof over the temporal fossa. The zygomatic arch is relatively small, and it is overhung by the base of the horn core. Behind the glenoid is an expansion of the squamosal, carrying the tube of the external auditory meatus, and a ridge connects this expanded! bone with the lambdoidal crest. Below the external auditory meatus is the paroccipital process, a fairly short but heavy structure. The occipital condyles are quite large, which would indicate that the cervical vertebrae were heavy. The supraoccipital is constricted above the foramen magnum, but it flares widely at the lambdoidal crest. The pits in the supraoccipital for the attachment of the semispinalis capitis muscle are shallow, as is characteristic of the Palaeotragine group of giraffids, as contrasted with the more advanced Sivatherines and Giraf- fines, in which these pits are quite deep. In a consideration of the basicranium it may be well to take up the several elements separately, in order that they may be discussed accord- ing to their positions and functions. Therefore, the floor of the brain case will first be considered, after which there will be a discussion of the auditory apparatus, and then finally some attention will be paid to the articulations and the attachments for the jaw and for the masticating muscles. The basicranium is rather wide in comparison to its length. The basioccipital is broad and is characterized by its large basilar tubercles, for the attachments of the rectus capitis ventralis major muscles. These tuberosities are much more prominent than in the modern okapi or giraffe; indeed, they are relatively larger than in Samotherium or Palaeo- tragu.s, and are quite comparable to the same structures in Bramatherium. The basisphenoid has low ridges along either edge. In contradistinction to most of the Giraffidae, the posterior nasal choanae are situated somewhat behind the third molar, instead of being directly in line with the posterior border of the last tooth. The pterygoids of Giraffokeryx are developed much as they are in the okapi. Owing to the fact that this animal was fully adult, the sutures are largely Sphenopalatine - ( foraimen

thmoidPt|erio ' foramen onasal 6 choonae

q Optic I 11 m\tc Glnodforamen / (hf \rch Foramen rotundum andforomenrIAK.U laceru anterius Glenoid Foramen ovalIe tchianand 3 fdramenla Internal carotid Postqlenoi rudierummeu Temp )ra foramen udi si I canal pit blla!'tubercleso ,K Externccipalf lj ) a-basioccipil L Ix u oucditorj mastoid(F \ meatus, \Joamen, posne r us Juqular PrOCC1pi4-V, pr css-.f ;'Condqljar exoccipital Xoromen

, /\\Pn

Fig. 4.-Diagram of basicranium and adjacent regions of the skull of Giraffokeryx punjabiensi8 Pilgrim. Names of foramina shown on left side of diagram; position of nerves and blood vessels on right.}\~~~ 8 1933] SKULL OF GIRAFFOKERYX PUNJABIENSIS 9 obliterated, making precise determinations of the basicranial elements uncertain, but undoubtedly the alisphenoid ran well up within the orbit. Coming now to the basicranial foramina, it may be seen that the ethmoid and optic foramina are large and separated from each other. As in the okapithere are two large foramina at the base of the alisphenoid, the anterior one of which would seem to be a fusion of the foramen rotundum and the foramen lacerum anterius, while the one posterior to it is undoubtedly the foramen ovale. Thus the anterior foramen served as a common passage for the oculomotor, the trochlear, the first and second branches of the trigeminus and the abducens nerves, while the posterior foramen transmitted the mandibular or third branch of the trigeminus. The development of the foramen lacerum medius is similar to that in the okapi, namely, it together with the opening for the Eustachian tube, is concealed above the anterior point of the bulla. The stylomastoid foramen is large, evidently carrying a well developed facialis nerve, and again, the postglenoid foramen is of unusually great size. In both of these characters, Giraffo- keryx is similar to the okapi. The foramen lacerum posterius is of normal development, and separate from the foramen lacerum medius. In Okapia there are two well developed condylar foramina for the branches of the hypoglossal nerve. It is difficult to be certain about this region in the skull of Giraffokeryx, because of the manner of preservation, but probably these foramina were well developed. The reader is referred to the accompanying figure of the basicranium of Giraffokeryx, in which the various foramina are labeled. The auditory bulla is of medium size, though not as large nor as in- flated as in Okapia. In Giraffokeryx the bony tube of the external auditory meatus points posteriorly, while in Okapia this structure is placed almost transversely. Coming now to the articulation of the lower jaw, we see that the glenoids are convex as in the okapi. The pterygoids are fairly heavy, showing the power of the internal pterygoid muscles to move the jaw from side to side. The short heavy paroccipital processes would point to strong digrasticus muscles. The palate is wide and flat. The cheek teeth are well worn, again an indication that this was a fully adult animal at the time of its death. There are but three premolars, the first one having been lost during the course of evolutionary development. The muzzle is missing anterior to P2, but undoubtedly there was no canine, and the premaxillaries were edentulous. The cheek teeth are brachyodont, showing this animal to be 10 AMERICAN MUSEUM NOVITA TES r [No. 632 primitive, even though aberrant, and they are marked by the characteristic giraffid rugose enamel. The second and third premolars are greater in length than in width, and each tooth is characterized by a strong anterior style, and an internal posterior swelling. In the last premolar the width considerably exceeds the length. The molars are approximately quadrate, with the protocone and metaconule of about, equal size, and with strong parastyles and mesostyles. An isolated tooth (Amer. Mus. No. 19611) shows the character of thle last premolar. In this tooth the median outer barrel is well developed and there is a swelling on the posterior border of the inner crescent, which when worn forms a small enamel lake within the dentine. Amner. Mus. No. 19472, listed above, demonstrates the form of the molars much more clearly than do the teeth in the skull. In each molar, along the ectoloph, the parastyle and mesostyle are strongly developed; the anterior barrel is strong, but on the posterior part the ectoloph is flat. A similar condition holds in Palaeotragus, but it is to be noted that the styles and barrels in Giraffokeryx are divergent, while in Palaeo- tragus they are parallel, a fact that finds its explanation in the greater hypsodonty, and consequently the more advanced state of the teeth, in the latter genus. The anterior inner crescent of each molar in Giraffokeryx has a slight cingulum. Moreover, this crescent extends lingually further than does the posterior crescent, which gives the anterior half of the tooth a greater transverse measurement than the posterior moiety. There are no inner pillars on the molars as in Palaeotragus. Pilgrim speaks of the absence of an enamel fold in the posterior crescent of the molar of Giraffokeryx. Such a fold is present to a greater or lesser extent in all of the American Museum specimens, and in some of them it forms an enamel loop within the dentine. All in all, the upper teeth of Giraffokeryx are very similar to those of Palaeotragus, demonstrating (as has been pointed out above, in connec- tion with certain cranial characters) the close relationship existing between these two genera.

THE MANDIBLE Amer. Mus. No. 19587, listed above, contains a complete set of grinding teeth, which differ from the teeth of Palaeotragus and Okapia mainly by virtue of their relative narrowness. The second premolar is relatively simple, consisting of a median cone or point, from which a single ridge extends forward, and to the back of which there is a posterior ridge, dividing to form a Y. The third premolar is more advanced, 1933] SKULL OF GIRAFFOKERYX PUNJABIENSIS 11 having the main cusp placed exteriorly. Anteriorly the ridge divides so as to form two prongs pointing lingually, while -posteriorly there are three such prongs. There is a low internal wall to this tooth. In the fourth premolar the internal wall has developed to form a high internal ridge, rising to a point opposite the external cone. The external ridge divides into three prongs posteriorly, as in the preceding tooth. In P3 and P4 there are anterior external cingula. The molars consist of the familiar artiodactyl crescents, and naturally in the third molar there is an additional posterior loop or talonid. Cingiala are present anteriorly on the internal and external sides. The enamel is

-~~~--;----s0 S~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~8'~~~~~~~~~~~~A.M.19587- - - .g Fig. 5.-Giraffokeryx punjabienis Pilgrim. Amer. Mus. No. 19587. Right ramus of mandible, lateral view. Crown view of grinding teeth, above. Re- stored portions indicated by dotted lines. The length of the restored premolar- canine diastema is based on comparisons with Palaeotragus. One-fourth natural size. rugose. There is a strong external median pillar in the first molar, and rudimentary knobs in the succeeding teeth. The incisor alveoli are present, but a portion of the ramus forming the diastema was lost. Undoubtedly this diastema was long, as in Palaeotragus or Okapia, and it would seem as if the ascending ramus was shaped much as in the latter genus. MEASUREMENTS SKULL.-Amer. Mus. No. 19475. Length, P2 to condyle 326 mm. Estimated total length 500 Height above MS (to superior border of orbit) 85 Restored height 110 Antero-posterior diameter of orbit 66 Postorbital length (front of orbit to condyle) 222 12 AMERICAN MUSEUM NOVITATES [No. 632 MEASUREMENTS (Cordinued) SKULL.-Amer. Mus. No. 19475 (Continued) Estimated preorbital length 278 Width between tips of anterior horn cores 277 Width between tips of posterior horn cores 403 Width of confluent base of anterior horn cores 142 Distance between bases of posterior horn cores 123 Greatest width across squamosals 144 Width at narrowest portion of parietals 56 Width across zygomatic arches 179 Width of maxillae above p2 80 Width of palate at M3 66 Width of occipital condyles 76 Distance, anterior border of foramen magnum to border of posterior nasal choanae 150 Width between orbits 134 Width between tips of paroccipital processes 88 MANDIBLE.-Amer. Mus. No. 19587 Depth of ramus below third molar 43 UPPER DENTITION.- Giraffokeryx punjabiensis Palaeotragus microdon Amer. Mus. No. 19475 (After Bohlin)I Antero-posterior Transverse Antero-posterior Transverse p2 22 mm. 19mm. 18 mm. mm. P3 20.5 20 18 20 P4 17.5 21 18 23 Ml 22 24 23 27 M2 25 27 26 30 M3 24.5 26 25 27 LOWER DENTITION. Amer. Mus. No. 19587 (After Bohlin)2 P2 18 9 10 P3 20.5 12 15 13 P4 24 15 18 15 Ml 24 16 17 M2 25 27 23 18 M3 37 17 36 18 Ratio, premolar to molar length Upper dentition 83 73 Lower dentition 71 558

'Bohlin, B. 1927. 'Die Familie Giraffidae.' Pal. Sinica, Ser. C, IV, Fas. i, p. 27, column d in table. 2Bohlin, B. 1927. op. cit., p. 29, column I in table. 3Missing measurements obtained from other specimens listed by Bohlin. A.M. 1947 5 A.M. 19587

Fig. 6.-Restoration of the skull and mandible of Giraffokeryx punjabiensis Pilgrim, showing the probable appearance of the skull with the crushing removed. One-fourth natural size.

13 14 AMIERICAN MUSEUM NOVITATES [NO. 632 BIBLIOGRAPHY BOHLIN, BIRGER. 1927. 'Die Familie Giraffidae.' Pal. Sinica, Ser. C, IV, Fas. 1. FORSYTH MAJOR, C. J. 1891. 'On the Fossil Remains of Species of the Family Girafifdae.' Proc. Zool. Soc. London, pp. 315-326. FRAIPONT, J. 1908. 'L'Okapi.-Ses affinit6s avec les Giraffid6s vivants et fossiles.' Acad. Roy. de Belgique. Bull. de la classe des sciences. No. 12, Bruxelles. LANKESTER, E. RAY. 1907. 'The Origin of the Lateral Horns of the Giraffe in Foetal Life on the Area of the Parietal Bones.' Proc. Zool. Soc. London, pp. 100-115. 1907. 'The Existence of Rudimentary Antlers in the Okapi.' Proc. Zool. Soc. London, pp. 126-135. 1910. 'Monograph of the Okapi.' London. MATTHEW, W. D. 1929. 'Critical Observations upon Siwalik Mammals.' Bull. Amer. Mus. Nat. Hist., LVI, Art. VII, pp. 535-554. MuRIE, J. 1872. 'On the Horns, Viscera and Muscles of the Giraffe.' Ann. Mag. Nat. Hist., 4th Series, IX, pp. 177-195. PILGRIM, G. E. 1910. 'Notices of New Mammalian Genera and Species from the Tertiaries of India.' Rec. Geol. Surv. India, XL, p. 69. 1910. 'Tertiary Freshwater Deposits of India,' Rec. Geol. Surv. India, XL, p. 203. 1911. 'The'Fossil Giraffidae of India.' Mem. Geol. Surv. India, Pal.- Indica, N.S. IV, No. 1. 1913. 'Correlation of the Siwaliks with Mammal Horizons of Europe.' Rec. Geol. Surv. India, XLIII, Pt. 4. WEBER, MAx. 1904. 'DieSaugetiere.' Jena. The following titles refer to genera belonging to the Protoceratidae. BARBOUR, E. H. 1905. 'Notice of a New Fossil Mammal from Sioux County, Nebraska.' Neb. Geol. Surv., II, Pt. 3, four pages, one plate. STIRTON, R. A. 1932. 'A New Genus of Artiodactyla from the Clarendon Lower Pliocene of Texas.' Univ. of California Publ., Bull. Dept. Geol. Sci., XXI, No. 6, pp. 147-168, Ple. vi-xi.